Ricin is a heterodimeric protein toxin of molecular weight 65,000. Its B chain is a lectin which binds to cell surfaces and facilitates uptake of the toxin. The A chain kills the receptor cell by enzymatically attacking its ribosomes and inhibiting protein synthesis. Although ricin kills tumor transformed cells at least 100 times faster than normal cells, the toxin has been used most effectively when combined with antibodies against tumor cells to form an immunotoxin. Such conjugates are now in clinical trails for special therapeutic cases. A molecular model based on x-ray diffraction is now underway with 85% of the backbone in place. We propose to complete the molecular model of ricin and to investigate the detailed structure of the toxin, particularly with respect to the cell surface binding area of the B chain. Whole animal studies suggest that this information can be used to design and engineer a B chain which will have increased relative specificity for tumor cells targeted by the antibodies and decreased non-specific toxicity. The gene for ricin has been cloned by Cetus Corporation and a collaboration has been formed with them to carry out this genetic engineering project. We have recently discovered that the single chain antiviral protein PAP is genetically related to ricin A chain. Since three dimensional structure is even more conservative than linear sequence, it therefore seems likely that PAP will resemble the A chain fold. We have crystals and 2.5A data for PAP and for one derivative. In addition to a continuing search for other good derivatives, we propose to use computer search methods to fit the A chain model to PAP density and to refine the structure. Chemical work will continue on both ricin-like and PAP-like toxins, including searches for active site labels and pseudosubstrates to be viewed by difference Fourier. Finally, we propose to continue our efforts to clone and sequence cDNA corresponding to PAP. This effort should provide us with the amino acid sequence of PAP and perhaps the ability to test mechanistic theories with site directed mutagenesis.